Helping older adults spot AI deepfakes
by Katy Keffer, staff writer
The use of AI-generated deepfakes in scams is increasing, particularly those targeting older adults who may not recognize fraudulent activity. Lakshmi M. Katrapati, a student pursuing a computer science degree at George Mason University, is working to change that. As technology advances rapidly, she reaches out to senior living communities and centers to give back to the community. Since March 2026, she has given nearly 50 specialized presentations on deepfakes in Fairfax and Loudoun counties and Washington, D.C. At an Oct. 2 event at The Kensington Reston, she explained the topic to an engaged group that asked probing questions and left better prepared to spot deepfake videos.
Katrapati’s presentation, “AI For You: Staying Safe and Creative,” introduced broad AI concepts through interactive ChatGPT prompts based on audience questions. She also highlighted recent case studies in which deepfakes were used to manipulate people for financial gain. She shared multiple examples of how to distinguish authentic videos from fake ones, including videos of former Canadian Prime Minister Justin Trudeau recommending a book and American businessman Bill Gates discussing technology.
But how can we recognize whether something is fake or real? That is the focus of Katrapati’s extracurricular presentations. She educates audiences about what to do if they are contacted by a stranger or believe an inquiry may not be legitimate.
If on the phone, in a Zoom meeting or on a FaceTime call, “look and listen for pauses in responses,” she recommends. These could indicate that AI is taking a moment to generate a response. During a video call, “you can have the person wave their hand in front of their face to verify authenticity.” She notes that visual inconsistencies can be a sign of a deepfake, although sophisticated deepfakes can be difficult to detect.
For potential voice scams, such as those targeting grandparents with a supposed grandchild’s voice, Katrapati suggests asking something personal but not sensitive that only the other person would know. Deepfakes can be created using actual video, images or audio, including material that may be available on a family member’s social media feeds. Recordings also can be created to mimic loved ones’ voices.
“Deepfakes are essentially another form of misinformation,” Katrapati said, “which has been around for some time.” As with other misinformation, “it’s always important to fact-check the information with reputable sources.” Government websites, nonprofit organizations and media outlets such as The Washington Post and The New York Times are among the sources to consider.
Audience members at The Kensington Reston also asked how AI can be used in daily life and how it could help in medicine and health care. “Its ability to summarize lots of material for the most relevant insight is beneficial,” Katrapati said. She also offered other tips for staying safe.
“Create a family safe code or password as one form of verification,” she advised. She cautioned people not to share private or intimate information that could be recorded and later used for fraudulent purposes. She also recommended letting a call go to voicemail if the caller’s identity is uncertain. If someone is concerned about a suspicious call, she advised hanging up and independently calling the person back using a known phone number rather than relying on the number provided by the caller.
The group asked how to stay ahead of AI’s advances, and Katrapati acknowledged that AI and deepfake technology are changing rapidly. As a junior at George Mason University, she hopes to work in the industry after completing her degree. Two options are top of mind: “I’d like to help make the AI models safer or even pursue the legislative side.”
Comments